[1]张平均,翁悦,王小红,等.基于改进UNet的人造板表面缺陷的图像分割方法[J].福建工程学院学报,2022,20(04):373-377.[doi:10.3969/j.issn.1672-4348.2022.04.011]
　ZHANG Pingjun,WENG Yue,WANG Xiaohong,et al.Image segmentation method of surface defects of wood-based panels based on improved UNet[J].Journal of FuJian University of Technology,2022,20(04):373-377.[doi:10.3969/j.issn.1672-4348.2022.04.011]
点击复制

基于改进UNet的人造板表面缺陷的图像分割方法()
分享到：

《福建工程学院学报》[ISSN:2097-3853/CN:35-1351/Z]

卷:
第20卷
期数:
2022年04期
页码:
373-377
栏目:
出版日期:
2022-08-25

文章信息/Info

Title:
Image segmentation method of surface defects of wood-based panels based on improved UNet
作者:
张平均; 翁悦; 王小红; 李稳稳; 林艺斌
福建工程学院电子电气与物理学院
Author(s):
ZHANG Pingjun;  WENG Yue;  WANG Xiaohong;  LI Wenwen;  LIN Yibin
School of Electronic, Electrical Engineering and Physics, Fujian University of Technology
关键词:
聚焦注意力机制模块; UNet; 残差网络; 人造板表面缺陷; 图像分割
Keywords:
attention focusing mechanism module;  UNet;  residual network;  surface defects of wood-based panels;  image segmentation
分类号:
TP391.4
DOI:
10.3969/j.issn.1672-4348.2022.04.011
文献标志码:
A
摘要:
为满足人造板表面缺陷图像分割的精度要求，提出了一种改进的UNet 语义分割网络模型。 在传统的UNet 网络结构上将编码部分改进为残差网络ResNet50 并去掉连接层与平均池化层，网络通过残差块堆叠获取更多特征的底层信息；同时在跳跃连接中嵌入聚焦注意力机制的模块，抑制干扰信息，保留有效位置信息，聚焦缺陷区域并加强学习。 对4 种UNet 网络模型的人造板表面缺陷图像分割进行仿真比较，结果表明，融合聚焦注意力机制的残差UNet 网络模型在像素准确率和平均交并比等指标上有较大提升，分割精度较高。
Abstract:
In order to meet the requirements of the precision of image segmentation of surface defects of wood-based panels, an improved UNet semantic segmentation network model was proposed. The coding part of the traditional UNet network was modified into residual network ResNet50, and the connection layer and average pooling layer were removed. The network was stacked with residual blocks to obtain more underlying information of features. At the same time, the module of attention focusing mechanism is embedded in the jump connection to suppress interference information, retain effective location information, focus defect location and enhance learning. The simulation comparison of image segmentation of surface defects of wood-based panels based on four UNet models shows that the residual UNet model integrating the attention focusing mechanism has been greatly improved in pixel accuracy and average intersection ratio, as well as the segmentation accuracy.

参考文献/References:

[1] 朱志, 丁相君. 浅析人造板砂光后的表面缺陷[J]. 农业与技术, 2013, 33（11）: 111.[2] 魏智锋, 肖书浩, 蒋国璋, 等. 基于深度学习的人造板表面缺陷检测研究[J]. 林产工业, 2021, 58（2）: 21-26.[3] 郭慧, 王霄, 刘传泽, 等. 基于灰度共生矩阵和分层聚类的刨花板表面图像缺陷提取方法[J]. 林业科学, 2018, 54（11）: 111-120.[4] 郭慧, 王霄, 刘传泽, 等. 人造板表面缺陷检测图像自适应快速阈值分割算法[J]. 林业科学, 2018, 54（11）: 134-142.[5] 杨传礼,张修庆.基于机器视觉和深度学习的材料缺陷检测应用综述[J/OL].材料导报,2022(16):1-19[2022-04-15].http:∥kns.cnki.net/kcms/detail/50.1078.TB.20210728.1056.009.html.[6] HE T, LIU Y, XU C Y, et al. A fully convolutional neural network for wood defect location and identification[J]. IEEE Access, 2019, 7: 123453-123462.[7] 宫艳晶, 黄民, 黄小龙. 基于改进ResNet-UNet的火焰图像分割方法[J]. 北京信息科技大学学报（自然科学版）, 2021, 36（5）: 39-44.[8] RAHMAN M R U, CHEN H Y. Defects inspection in polycrystalline solar cells electroluminescence images using deep learning[J]. IEEE Access, 2020, 8: 40547-40558.[9] 谢舰, 姚剑敏, 严群, 等. 基于深度学习的磁瓦表面缺陷分割与识别[J]. 液晶与显示, 2021, 36（5）: 713-722.[10] 徐昭洪, 刘宇, 全吉成, 等. 基于VGG16预编码的遥感图像建筑物语义分割[J]. 科学技术与工程, 2019, 19（17）: 250-255.[11] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]∥2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016: 770-778.[12] OKTAY O, SCHLEMPER J, FOLGOC L L, et al. Attention U-net: learning where to look for the pancreas[EB/OL]. 2018: arXiv: 1804.03999[cs.CV]. https://arxiv.org/abs/1804.03999.

更新日期/Last Update: 2022-08-25